Stakeholder Perspectives on the Draft Implementation Act for Joint Clinical Assessments: Key Insights and Concerns

OBJECTIVES: The draft Implementation Act (IA) for Joint Clinical Assessment (JCA) of medicinal products was released for public commentary within the framework of the Health Technology Assessment Regulation (HTAR) on March 5, 2024. This study aims to classify & analyze responses received from the public consultation using Artificial Intelligence (AI) and human review.

METHODS: A total of 129 responses were collected by the April 2, 2024 deadline. GPT-4 was used to extract structured data and conduct thematic analysis. Human review was used to verify the AI findings. The primary issues raised by Company/Business stakeholders were identified and examined using AI followed by human verification.

RESULTS: Received response on the draft IA JCA were categorized by organizational type: Company/Business, Business Associations, Academic/Research Institutions, Non-governmental Organizations (NGOs), EU/Non-EU Citizens, Trade, Public Authorities, Environmental Organizations, and Others. Company/Business (21%) and Business Associations (19%) collectively composed 40% of the responses. The predominant concern (89%) from Company/Business stakeholders was the insufficient time allocated for high-quality dossier preparation. Additional significant issues included the inclusion of health technology developers (HTDs) throughout the JCA process especially during the scoping phase (70%), transparency on the overall JCA process (52%), and confidentiality matters (48%). The updated draft IA document incorporated several adjustments based on these feedback but fell short of fully addressing all concerns. AI and human extracted content aligned with minor differences.

CONCLUSIONS: The initial IA for JCA presented a number of issues for stakeholders that were partially addressed but will still lead to challenges related to increased demand for evidence, strain on company resources, and uncertainties in application. This study demonstrates that using AI versus human review yielded similar results. However, the time to analyze the data by AI was less than one tenth the time of human review.